S

Oip

TLa

Sb

a

ARAA

KTMcCPP

1

ttsasbPbed

iLPh

1d

Journal of Chromatography B, 879 (2011) 804–810

Contents lists available at ScienceDirect

Journal of Chromatography B

journa l homepage: www.e lsev ier .com/ locate /chromb

hort communication

n-line combination of monolithic immobilized pH gradient-based capillarysoelectric focusing and capillary zone electrophoresis via a partially etchedorous interface for protein analysis

ingting Wanga,b, Junfeng Maa, Shuaibin Wua,b, Liangliang Suna,b, Huiming Yuana,ihua Zhanga,∗, Zhen Lianga, Yukui Zhanga

Key Laboratory of Separation Science for Analytical Chemistry, National Chromatographic Research and Analysis Center, Dalian Institute of Chemical Physics, Chinese Academy ofciences, 457 Zhongshan Road, Dalian 116023, ChinaGraduate University of Chinese Academy of Sciences, 19A Yuquan Road, Beijing 100049, China

r t i c l e i n f o

rticle history:eceived 27 August 2010ccepted 10 February 2011vailable online 21 February 2011

eywords:wo-dimensional capillary electrophoresisonolithic immobilized pH gradient-based

a b s t r a c t

An integrated platform consisting of monolithic immobilized pH gradient-based capillary isoelectricfocusing (M-IPG CIEF) and capillary zone electrophoresis (CZE) coupled by a partially etched porousinterface was established. Since carrier ampholytes (CAs) were immobilized on monolith in M-IPG CIEFto form a stable pH gradient, subsequent depletion of CAs at the interface to prevent the interference onCZE separation and detection were avoided. Moreover, a partially etched porous capillary column, whichwas facile for fabrication and durable for operation, was exploited as the interface to combine M-IPG CIEFand CZE. The RSD values in terms of the migration time for M-IPG CIEF separation, transfer protein from

apillary isoelectric focusingapillary zone electrophoresisartially etched porous interfacerotein

the first dimension to the second dimension, and CZE separation, were 2.4%, 3.9% and 2.3%, respectively.With a 6-protein mixture as the sample, two-dimensional capillary electrophoresis (2D-CE) separationwas successfully completed within 116 min, yielding a peak capacity of ∼200 even with minute sampleamount down to 5.0 �g/mL. The limit of detection was 0.2 �g/mL. In addition, proteins extracted frommilk were used to test the performance of such a 2D-CE separation platform. We expect that such a novel2D-CE system would provide a promising tool for protein separation with high throughput and high peak

capacity.

. Introduction

Efficient separation is a prerequisite step for the characteriza-ion of proteins, especially the complex protein mixtures. Althoughwo-dimensional polyacrylamide gel electrophoresis (2D-PAGE) istill a widely recognized approach, it suffers from unavoidable dis-dvantages, such as time and labor consuming, poor capacity toeparate large and hydrophobic proteins, and extremely acidic andasic proteins [1,2]. However, many of the disadvantages of 2D-

AGE are not shared by two-dimensional (2D) separation methodsased on capillary electrophoresis (CE). Moreover, high columnfficiency, high resolution, and ease in coupling with a variety ofetectors (e.g. laser-induced fluorescence and mass spectrome-

Abbreviations: M-IPG CIEF, monolithic immobilized pH gradient-based capillarysoelectric focusing; CZE, capillary zone electrophoresis; CAs, carrier ampholytes;PAA, linear polyacrylamide; �-MAPS, 3-(trimethoxysilyl)propyl methacrylate;EG, poly(ethylene glycol); 2D-CE, two-dimensional capillary electrophoresis; HF,ydrofluoric acid.∗ Corresponding author. Tel.: +86 411 84379720; fax: +86 411 84379779.

E-mail address: lihuazhang@dicp.ac.cn (L. Zhang).

570-0232/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.jchromb.2011.02.020

© 2011 Elsevier B.V. All rights reserved.

try) are often demonstrated, thus 2D-CE separation systems havedrawn much attention recently [3–5].

Developing a reliable and robust interface is a key issue to con-struct an efficient 2D-CE separation system. Over the past years, avariety of 2D-CE interfaces have been introduced, including crossinterface [6], valve [7], microdialysis device [8–10] and wholeetched porous interface [11]. Dovichi’s group [6] first reporteda system for automated protein analysis using an interface thataligned two separating capillaries and two waste capillaries whichwere held in place by tightening the ferrules in the Valco cross.Rassi’s group [7] coupled capillary isoelectric focusing (CIEF) to cap-illary electrochromatography via a nanoinjector valve to perform2D separation. Mohan et al. [8] employed microdialysis junction asthe interface for on-line coupling of CIEF with transient capillaryisotachophoresis/zone electrophoresis (CITP-CZE) for the separa-tion of tryptic digests of proteins. In our group, microdialysis

junction was also developed as 2D-CE interface to achieve the bufferexchanging and necessary electrical connection [9,10]. With suchinterfaces, CIEF-capillary gel electrophoresis (CIEF-CGE) [9] andCIEF-capillary non-gel sieving electrophoresis (CIEF-CNGSE) [10]were successfully established. Besides the microdialysis devices, a

atogr. B 879 (2011) 804–810 805

wt

thadaiboaac

psawfwfetrm

2

2

afppmtlw(GwpAR84(L

2

mfive(

2

Cmw

T. Wang et al. / J. Chrom

hole etched porous capillary was developed as interface as well,hough some fragileness was observed [11].

CIEF is preferably used as the first dimension in the construc-ion of 2D-CE systems [7–11], due to its high separation resolution,igh peak capacity, and excellent concentration capability. Carriermpholytes (CAs) are indispensable to establish a stable pH gra-ient and usually reduce UV detection sensitivity due to the highbsorbance at low wavelengths. To solve this problem, monolithicmmobilized pH gradient (M-IPG) columns, by which CAs wereonded to the surface of the monolith in a capillary, were devel-ped in our group [12,13]. The M-IPG columns showed severaldvantages including the elimination of interference of CAs for sep-ration and detection, reduced diffusion of focused zones and easilyoupled with separation and/or detection systems.

Based on our previous work on the preparation of a whole etchedorous interface [11] and the construction of CIEF based separationystems [9–13], a partially etched porous capillary interface andnovel 2D-CE platform were developed herein. In this platform,ith the partially etched porous capillary column as the inter-

ace, for the first time, M-IPG CIEF was successfully hyphenatedith CZE for protein separation. The partially etched porous inter-

ace, which was prepared in a simple manner, illustrated effectivelectrical contact and improved robustness. The performance ofhe 2D-M-IPG CIEF-CZE platform was demonstrated by the sepa-ation of 6-standard protein mixture and proteins extracted fromilk.

. Materials and methods

.1. Materials

N,N,N′,N′-Tetramethylethylenediamine, ammonium persulfate,crylamide and N,N′-methylenebisacrylamide were purchasedrom Acros Organics (Geel, Belgium). Glycidyl methacrylate wasurchased from Fluka (St. Gallen, Switzerland). Ampholine (A5174,H 3.5–10.0), 3-(trimethoxysilyl)propyl methacrylate (�-MAPS),yoglobin from equine skeletal muscle (M1882, pI 7.3, 17.6 kDa),

rypsin inhibitor from soybean (T9003, pI 4.5, 23.0 kDa), �-actoglobulin A and B from bovine milk (L3908, pI 5.2, the molecular

eight of the dimer, 35.0 kDa) and ribonuclease A from pancreasR5500, pI 9.5, 13.5 kDa) were purchased from Sigma (Steinheim,ermany). Urease from sword bean (135143, pI 5.1, 480.0 kDa)as purchased from Merck (Darmstadt, Germany). Formamide wasurchased from Tianjin Bodi Chemical Corporation (Tianjin, China).zobisisobutyronitrile was purchased from The Fourth Shanghaiegent Plant (Shanghai, China). Poly(ethylene glycol) (PEG, MW000, 10,000) and albumin from bovine serum (RCCH735108, pI.9, 66.0 kDa) were from Sino-American Biotechnology CompanyShanghai, China). Skimmed milk was purchased from Mengniu Co.td. (Huhehaote, China).

.2. Sample preparation

Skimmed milk was centrifuged (AllegraTM 64R Centrifuge, Beck-an Coulter, Brea, CA, USA) at 15,000 × g for 30 min at 4 ◦C and

ltered by 0.45 �m filter to remove the solid. Then it was dried byacuum rotary evaporators (SPD131DDA-230, Thermo Fisher Sci-ntific, San Jose, CA, USA), and redissolved in 10 mmol/L Tris–HClpH 8.0) with the final concentration of 0.046 mg/mL.

.3. Preparation of M-IPG column

The inner surface of a silica capillary (Sino Sumtech, Handan,hina) was pretreated as described previously [13] with minorodifications. Briefly, the capillary was washed by 0.5 mol/L HCl,ater, 0.5 mol/L NaOH, water, and methanol, respectively, for

Fig. 1. Schematic diagrams of the partially etched porous interface (A), first dimen-sional separation (B) and second dimensional separation (C) of 2D-M-IPG CIEF-CZEsystem.

30 min at the flow rate of 2.0 �L/min, and dried with N2 at 70 ◦C for1 h. A solution of �-MAPS (50%, v/v in methanol) was injected intothe capillary for 10 min at the flow rate of 2.0 �L/min and kept atroom temperature for 24 h. Unreacted �-MAPS was washed withmethanol, and the capillary was purged with N2 at 70 ◦C for 1 h.20.6 mg glycidyl methacrylate and 34.3 mg CAs were dissolved in679.4 mg formamide. After vortexed for a few minutes, the mix-ture was placed in a water bath at 40 ◦C for 1 h, and then putin a refrigerator at 4 ◦C for 10 min. Subsequently, 45.0 mg PEG-8000, 25.0 mg PEG-10,000, 20.0 mg N,N′-methylenebisacrylamide,10.0 mg acrylamide and 0.5 mg azobisisobutyronitrile were addedinto the mixture. The mixture was vortexed and degassed with N2for 5 min, followed by injection into the �-MAPS coated capillary(100-�m i.d. × 375-�m o.d.) at the flow rate of 2.0 �L/min. Afterthat, with 20 mmol/L glutamic acid as anolyte buffer and 20 mmol/LNaOH as catholyte buffer, 400 V/cm was applied on the capillaryfor focusing the polymerization solution for 6 min. With both ends

sealed, the capillary was put in an oven at 70 ◦C for 20 h to immobi-lize pH gradient onto the monolith. Finally, the monolithic columnwas washed with methanol for 2 h, followed by washing with waterfor 1 h.

806 T. Wang et al. / J. Chromatogr. B 879 (2011) 804–810

F face. (b (C) cr

2

2w

ig. 2. Scanning electron micrographs of the partly etched fused-silica porous interlade before etching (150×); (B) outer surface of the interface after etching (150×);

.4. M-IPG CIEF separation

Electrophoresis experiment was performed on the TriSep-010GV (Unimicro Technologies, Pleasanton, CA, USA) equippedith a Data Module UV–visible detector and a high-voltage power

A) Outer surface of the fused-silica capillary after removing polyimide coating by aoss-section view of the interface after etching (500×).

supply. Workstation Echrom98 of Dalian Elite Analytical Instru-ment Co. Ltd. (Dalian, China) was used for data acquisition andanalysis.

M-IPG CIEF was performed in a 20 cm long M-IPG column. Cath-ode buffer was 20 mmol/L NaOH, and anode buffer was 20 mmol/L

T. Wang et al. / J. Chromatogr.

Fig. 3. Evaluation of the size of pores in the partially etched porous interface. (A)CZE coupled with the interface which was filled with sample; (B) CZE separationwithout the interface. Experimental conditions: sample, 2 mg/mL ribonuclease Ad�d

gwtcptaap

2

pw(cta2w

2

s0cwfsracfint

oaJg

posed to be in the low-nanometer range [11], below the resolution

issolved in 100 mmol/L Na2HPO4 (pH 8.0); non-coated capillary, 50-�m i.d. × 375-m o.d., 30 cm total length, 20 cm effective length; electric field strength, 333 V/cm;etection, 214 nm; buffer, 100 mmol/L Na2HPO4 (pH 8.0).

lutamic acid. The sample was injected until the whole capillaryas full. Then the voltage of 14 kV was applied for focusing. Once

he current reduced to ∼10% of the original value, the focusing wasonsidered to be complete. After that, the M-IPG column was cou-led with a capillary which made a detection window (10.0 cm ofotal length, 2.5 cm of effective length, 50-�m i.d. × 375-�m o.d.),nd then the zones were pumped through the detection window atflow rate of 70 nL/min by syringe pump (Baoding Longer precisionump Co. Ltd., Baoding, China).

.5. Preparation of linear polyacrylamide coating of CZE column

The inner wall of a CZE column was coated with linearolyacrylamide (LPAA) by the previously described method [14]ith minor modifications. Briefly, the �-MAPS coated capillary

50-�m i.d. × 375-�m o.d.) was flushed with degassed solutionontaining 25.0 mg acrylamide in 0.5 mL of water, 8 �L of N,N,N′,N′-etramethylethylenediamine (10%, v/v in water), and 8 �L ofmmonium persulfate (10%, w/v in water) at the flow rate of.0 �L/min and kept at room temperature for 12 h. The capillaryas ready to use after the residuals were rinsed out.

.6. Fabrication of the partially etched porous interface

On a Plexiglas reservoir, two holes were drilled straight, ashown in Fig. 1A. One fused silica capillary, on which about.5–2 mm length and 1/6 to 1/2 of diameter width of polyimideoating was removed by a blade, went through the holes andas screwed to the reservoir. HF was used to etch the exposed

used-silica section according to our previous procedure [11], withlight modifications. The fused-silica section which was partiallyemoved of polyimide coating was immersed in 40% HF for ∼1 ht room temperature in a well-ventilated hood. The etching pro-edure was monitored by measuring the current of the capillarylled with 100 mmol/L NH4HCO3 buffer periodically, and termi-ated until constant electrical conduction was established throughhe etched section wall.

Surface scanning and cross-sectional scanning electron images

f capillaries before and after HF etching were photographed withJEOL JSM-6360LV scanning electron microscope (JEOL, Tokyo,

apan) operated at 40 kV after coating the capillary segments withold–palladium in a vacuum evaporator.

B 879 (2011) 804–810 807

2.7. Construction of 2D-M-IPG CIEF-CZE platform

The partially etched porous interface was exploited for on-linecombination of the M-IPG CIEF and CZE separation system. M-IPGcolumn of 20 cm length acting as the first dimension was con-nected with the interface via a capillary of 2.5 cm length (50-�mi.d. × 375-�m o.d.) coated with LPAA, and a CZE capillary of 30 cmcoated with LPAA was served as the second dimension (Fig. 1B).The sample was introduced into the M-IPG column by manuallyusing a syringe. During M-IPG CIEF focusing, one platinum wire wasinserted into inlet reservoir, serving as the cathode, and anotherone was inserted into the interface, serving as the anode. Duringthe second dimensional separation, the outlet of CZE capillary wasserved as the cathode, and the interface was served as the anode(Fig. 1C). After completing M-IPG CIEF separation, the inlet of M-IPGcolumn was removed from the inlet vial, and connected to a syringepump. A fraction from the M-IPG column was pumped to the CZEcapillary at the flow rate of 70 nL/min for 1 min. With the pumpstopped, this fraction was further separated by CZE in a 10 min-run, and on-line detected with a window 20 cm from the partiallyetched porous junction. After that, the CZE separation was stopped,and the syringe pump connected to the M-IPG CIEF column wasrestarted for another 1 min fraction transfer, followed by CZE sep-aration. Such procedure was repeated until no more peaks wereobserved in CZE with all fractions from M-IPG CIEF analyzed.

3. Results and discussion

3.1. Characterization of the partially etched porous interface

The HF etched porous interface had been used in 2D-CE sys-tem [11], on-line concentration of proteins and peptides in CE [15],the interface of CE-MS [16,17] and with electrochemical detectionfor isolating the electrochemical detector from the CE electric field[18]. Capillary interfaces made by above methods were easy tobreak from the etched section. Zhang et al. [19] used a hole-openedcapillary for in-capillary SPE-CE concentration of chlorophenols. Inthis study, the idea of hole-opened capillary was applied with minormodifications, with the polyimide coating wall of capillary partiallyremoved. Furthermore, such a capillary was partially etched by HF.With this approach, the robustness of the capillary was significantlyenhanced.

The microstructure of the partially etched junction was accom-plished by scanning electron microscope. It could be seen thatbefore etching, the thickness of the removed polyimide coatingwall was ∼10 �m and the length of the removed polyimide coatingwall was ∼100 �m (Fig. 2A). After the treatment, the outer diame-ter of the capillary was distinctly decreased to ∼200 �m (Fig. 2B).When the fused-silica capillary was etched by the completely etch-ing method, as shown in our previous work [11], the thickness ofthe whole etched interface was about 11.0–18.0 �m, making theinterface easy to break. However, when the fused-silica capillarywas etched by the partially etching method, the thickness of etchedpart, less than 30% of the interface cross-section, as shown in Fig. 2C,was ∼10 �m, while that of the majority of the capillary wall was∼100 �m, rendering enough robustness of the interface for oper-ation. Moreover, it was found that the partially etched interfacecould be used several months without any observable deteriora-tion.

It should be noted that the diameter of the etched pores was sup-

limit of the electron microscope. Under the same conditions, with14 kV voltage, a 30 cm-length separation capillary and 20 mmol/Lglutamic acid as buffer, the current for 2D-CE was 2.0–2.2 �A, quiteclose to that measured in CZE without the interface (2.2 �A), indi-

808 T. Wang et al. / J. Chromatogr. B 879 (2011) 804–810

Fig. 4. M-IPG CIEF (A) and 2D-M-IPG CIEF-CZE (B) separation of 6-protein mixture. (Note: in (B) the remarked letters (3, 4, 6, 10) correspond to the M-IPG CIEF fractionsresolved by CZE.) Experimental conditions: for M-IPG CIEF separation: M-IPG column, 20 cm column length, 100-�m i.d. × 375-�m o.d.; anolyte, 20 mmol/L glutamic acid,catholyte, 20 mmol/L NaOH; electric field strength, 622 V/cm; detection, UV, 214 nm; the flow rate of syringe pump, 70 nL/min; for CZE separation: capillary coated withL c fields nsiond L. Ide( .0 kDa

caatoaeosA

PAA, 50-�m i.d. × 375-�m o.d., 30 cm total length, 20 cm effective length; electriystem: the separated proteins from first dimensional were injected to second dimeissolved in 10 mmol/L Tris–HCl (pH 8.0), concentration of each protein was 5.0 �g/mpI 5.1, 480.0 kDa); �-lactoglobulin (pI 5.2, the molecular weight of the dimmer, 35

ating that the pores of this porous structure were large enough tollow the permeation of small electrolyte ions upon application ofpotential to the system. A linear relationship (R2 = 0.999) between

he current in the separation capillary and the applied voltage wasbtained, suggesting that constant electric conductivity in the sep-

ration capillary could be achieved. The resistance of the partiallytched joint was calculated to be 0.71 × 109 �, which was the halff whole etched joint [20]. To further investigate the effect of poresize on the penetration of proteins, a small protein ribonuclease(13.5 kDa) was chosen as the sample. According to the previous

strength, 467 V/cm; buffer, 20 mmol/L glutamic acid; for the 2D-M-IPG CIEF-CZEal by a syringe pump, 70 nL/min; the injected time of every fraction, 1 min; sample,ntified peaks: trypsin inhibitor (pI 4.5, 23.0 kDa); albumin (pI 4.9, 66.0 kDa); urease

); myoglobin (pI 7.3, 17.6 kDa); ribonuclease A (pI 9.5, 13.5 kDa).

method to evaluate pore size on etched capillaries [21], as it canbe seen from Fig. 3, when ribonuclease A (13.5 kDa), the smallestprotein applied in this study, was added in the inlet vial, and con-tinuously injected by EOF, the protein signal was observed from3 min (line (B)). While with ribonuclease A added in the interface

buffer, and continuously injected by EOF under the same conditions(line (A)), no protein signal could be observed, which indicated thatthe pores in the wall could restrict ribonuclease A from passingthrough the interface, ensuring no protein loss during 2D separa-tion. The adsorption of ribonuclease A on uncoated silica capillary

T. Wang et al. / J. Chromatogr. B 879 (2011) 804–810 809

F milk.1

mtov

3

tgpgltNca6refaocwChC

iaermtfmbfotmiT�

ig. 5. 2D-M-IPG CIEF-CZE separation electropherogram of proteins extracted from.5 min. Other experimental conditions were the same as in Fig. 4.

ight occur. However, with high concentration protein, the adsorp-ion could not affect the evaluation of pore size. Therefore, the poresf the porous junction could accommodate protein molecules, pre-enting them from permeating the porous junction.

.2. Evaluation of 2D-M-IPG CIEF-CZE system

Due to the existence of back pressure, the migration time andhe peak width of proteins upon pressure-driven mobilization wasreatly influenced by the length of M-IPG column when the syringeump was employed. With the mixture of trypsin inhibitor, myo-lobin, and ribonuclease A as the sample, the effect of columnength on the migration time and peak width of the last eluted pro-ein, ribonuclease A, was studied. By comparison, taken 20 mmol/LaOH as catholyte and 20 mmol/L glutamic acid as anolyte, with theolumn length, respectively, as 40, 30 and 20 cm, the migration timend the peak width of ribonuclease A were decreased from 100.2,1.7 to 18.8 min, and 8.4, 8.1 to 1.1 min in the M-IPG CIEF separation,espectively. Therefore, a 20 cm long M-IPG column was applied tostablish the 2D-M-IPG CIEF-CZE platform. Furthermore, to per-orm 2D separation, the buffer of CZE should be the same as thenode buffer of M-IPG CIEF. Although the traditional anode bufferf CIEF was composed of phosphoric acid and acetic acid, the highurrent generated in CZE separation might interrupt 2D separationith bubbles formed. Therefore, glutamic acid was chosen as theZE separation buffer, beneficial to generate low current, even withigh voltage applied. Moreover, to achieve high speed analysis inZE, the optimal applied electric field strength was 467 V/cm.

The mixture of myoglobin and ribonuclease A was used to exam-ne the reproducibility of M-IPG columns. Under manual samplingnd mobilization by syringe pump, for one analysis performed inach of the three parallel columns, the RSD value in terms of theesolution of two proteins was 1.5% and the RSD values for theigration time of these two proteins were 2.4% and 0.8%, respec-

ively, demonstrating the good reproducibility of such columnsor M-IPG CIEF separation. The RSD values for the peak area of

yoglobin and ribonuclease A were 23.5% and 6.0%, which shoulde further improved. However, if two consecutive runs were per-ormed on the same M-IPG column, the RSD value for the resolutionf these two proteins was 7.4%, and the RSD values for the migra-

ion time of two proteins were 14.6% and 13.5%, respectively which

ight be caused by the possible collapse of the monolithic matrix,ndicating that it was better to use the M-IPG column only once.he RSD values for the migration time and peak area to transfer-lactoglobulin from the first dimension to the second dimension

Experimental conditions: sample, 0.046 mg/mL; the injected time of every fraction,

in ten consecutive runs were 3.9% and 17.4%, respectively. In CZEseparation, the RSD values for the migration time and peak areaof ribonuclease A were 2.3% and 7.2% (n = 21), respectively, simi-lar to that reported in the previous work [22]. The reproducibilityof M-IPG CIEF separation, protein transferring and CZE separationwere evaluated, respectively, and the acceptable RSD values couldbe obtained, indicating that the 2D-M-IPG CIEF-CZE system alsohad good reproducibility.

A 6-protein mixture with 5.0 �g/mL of each protein was used todemonstrate the utility and the resolving ability of the developed2D-M-IPG CIEF-CZE system. Fig. 4A presents the M-IPG CIEF sep-aration of the 6-protein mixture. Three proteins, trypsin inhibitor,albumin and ribonuclease A were baseline resolved. However, theother 3 proteins, urease, �-lactoglobulin, and myoglobin, wereeluted together. A total of 10 fractions (1, 2, . . ., 10) from the M-IPG column were injected into CZE, and 4 fractions were recordedwith obvious UV signals, as shown in Fig. 4B. Fractions of 3, 4 and10 yielded one peak, respectively, corresponding to the three sep-arated proteins by M-IPG CIEF. The co-eluted fraction containing3-protein mixture from M-IPG CIEF yielded baseline separationby CZE (Fig. 4B6). The elution order of proteins was determinedaccording to the charge/size ratio of proteins, as supposed bySheng et al. [23]. Since the pI of myoglobin is the highest, and themolecular weight is the lowest among the three proteins, with thehighest charge/size ratio in acidic buffer, it was the first protein tobe detected in CZE, followed by �-lactoglobulin and urease, withdecreased charge/size ratios. The limit of detection was 0.2 �g/mL,which was calculated at S/N > 3 by ribonuclease A, with the mediumintensity of UV signal among the 6-protein mixture. Therefore, effi-cient analysis of minute samples could be anticipated with such a2D-CE system.

The peak capacity was calculated according to the criteria pro-posed by 2D separation [24]. The peaks in CZE separation elutedover a range of 5.30 min, from 2.00 min to 7.30 min, and the peakwidth (4�) in CZE separation was 4.0 s. Thus the peak capacity ofCZE was ∼20. Since 10 fractions were eluted from M-IPG CIEF ontoCZE, the peak capacity was ∼200.

3.3. Application

Proteins extracted from cow milk, which mainly composed ofcaseins (pI 4.5, 80%), �-lactoglobulin (pI 5.2, ∼10%), �-lactalbumin(pI 4.2–4.5, ∼4%), albumin (pI 4.9, ∼1%) and other proteins (∼5%)[25], were used to evaluate the performance of the 2D-M-IPG CIEF-CZE system. A total of 8 fractions from the M-IPG column were

8 atogr.

i(tdwpcwtmie

ipflMmc

4

IpvtffCti

A

Cgfi

[

[

[

[

[

[[

[[

[[[[22] T.T. Wang, J.F. Ma, G.J. Zhu, Y.C. Shan, Z. Liang, L.H. Zhang, Y.K. Zhang, J. Sep. Sci.

33 (2010) 3194.

10 T. Wang et al. / J. Chrom

njected into CZE, and 4 fractions yielded detectable UV signalsFig. 5). In the representation, M-IPG CIEF cycle number was alonghe x-axis, and CZE separation time was along the y-axis, and theensity at each point is proportional to the UV signal. Many spotsere observed in the early 4 eluted fractions, which were in theosition of acidic end of M-IPG column, in accordance with theomposition of milk. As shown in the inset of Fig. 5, five main peaksere resolved in the first fraction, and the peaks a and b were iden-

ified as albumin and �-lactoglobulin, respectively, according toigration time of each protein in the standard proteins separation

n 2D-M-IPG CIEF-CZE system, showing the excellent separationfficiency of 2D-CE.

In the 2D-M-IPG CIEF-CZE system, overall resolution was greatlymproved, although might be not as impressive as some otherlatforms concerning 2D-CE separation followed by laser induceduorescence detection [4]. It is anticipated that the constructed 2D--IPG CIEF-CZE platform might be applicable to the separation ofore complex samples (e.g. proteomic samples from body fluids,

ells and tissues).

. Concluding remarks

In this work, the feasibility of an on-line combination of M-PG CIEF and CZE system for protein analysis by a partially etchedorous interface has been demonstrated. Compared with the pre-iously reported 2D-CE using the whole etched porous interface,he partially etched interface showed better robustness and easierabrication. Both standard proteins mixture and proteins extractedrom milk were successfully separated. Therefore, the 2D-M-IPGIEF-CZE system might provide a powerful tool for protein separa-ion. Further work on coupling this system with mass spectrometrys undergoing for the top-down profiling of proteomic samples.

cknowledgements

We are grateful to Prof. Guowang Xu (Dalian Institute ofhemical Physics) for the critical reading and valuable sug-estions on this manuscript. We are also grateful for thenancial support from National Basic Research Program of China

[[

[

B 879 (2011) 804–810

(2007CB914100), National Natural Science Foundation (20775080and 20935004), Creative Research Group Project by NSFC(21021004), Knowledge Innovation Program of Chinese Academyof Sciences (KJCX2YW.H09) and Sino-German Cooperation Project(GZ 3164).

References

[1] H.J. Issaq, T.D. Veenstra, BioTechniques 44 (2008) 697.[2] G. Candiano, L. Santucci, A. Petretto, M. Bruschi, V. Dimuccio, A. Urbani, S.

Bagnasco, G.M. Ghiggeri, J. Proteomics 73 (2010) 829.[3] D. Mohan, L. Pasa-Tolic, C.D. Masselon, N. Tolic, B. Bogdanov, K.K. Hixson, R.D.

Smith, C.S. Lee, Anal. Chem. 75 (2003) 4432.[4] S. Hu, D.A. Michels, M.A. Fazal, C. Ratisoontorn, M.L. Cunningham, N.J. Dovichi,

Anal. Chem. 76 (2004) 4044.[5] B.R. Fonslow, J.R. Yates III, J. Sep. Sci. 32 (2009) 1175.[6] D.A. Michels, S. Hu, R.M. Schoenherr, M. Regine, M.J. Eggertson, N.J. Dovichi,

Mol. Cell. Proteomics 1 (2002) 69.[7] M.Q. Zhang, Z.E. Rassi, J. Proteome Res. 5 (2006) 2001.[8] D. Mohan, C.S. Lee, Electrophoresis 23 (2002) 3160.[9] C. Yang, H.C. Liu, Q. Yang, L.Y. Zhang, W.B. Zhang, Y.K. Zhang, Anal. Chem. 75

(2003) 215.10] H.C. Liu, C. Yang, Q. Yang, W.B. Zhang, Y.K. Zhang, J. Chromatogr. B 817 (2005)

119.11] H.C. Liu, L.H. Zhang, G.J. Zhu, W.B. Zhang, Y.K. Zhang, Anal. Chem. 76 (2004)

6506.12] C. Yang, G.J. Zhu, L.H. Zhang, W.B. Zhang, Y.K. Zhang, Electrophoresis 25 (2004)

1729.13] G.J. Zhu, H.M. Yuan, P. Zhao, L.H. Zhang, W.B. Zhang, Y.K. Zhang, Electrophoresis

27 (2006) 3578.14] D. Schmalzing, C.A. Piggee, F. Foret, E. Carrilho, B.L. Karger, J. Chromatogr. A 652

(1993) 149.15] W. Wei, E.S. Yeung, Anal. Chem. 74 (2002) 3899.16] G.M. Janini, M. Zhou, L.R. Yu, J. Blonder, M. Gignac, T.P. Conrads, H.J. Issaq, T.D.

Veenstra, Anal. Chem. 75 (2003) 5984.17] J.T. Whitt, M. Moini, Anal. Chem. 75 (2003) 2188.18] X.B. Yin, H.B. Qiu, X.H. Sun, J.L. Yan, J.F. Liu, E.K. Wang, Anal. Chem. 76 (2004)

3846.19] L.H. Zhang, X.Z. Wu, Anal. Chem. 79 (2007) 2562.20] S. Hu, Z.L. Wang, P.B. Li, J.K. Cheng, Anal. Chem. 69 (1997) 264.21] X.Z. Wu, L.H. Zhang, K. Onoda, Electrophoresis 26 (2005) 563.

23] L. Sheng, J. Pawliszyn, Analyst 127 (2002) 1159.24] D.A. Wolters, M.P. Washburn, J.R. Yates III, Anal. Chem. 73 (2001)

5683.25] http://www.wheyoflife.org/faq.cfm.